OPULENT VOICE AND HAIFURAIYA, AN OPEN-SOURCE TECHNICAL BASELINE FOR THE NEXT GEOSTATIONARY AMATEUR RADIO PAYLOAD
A white paper for presentation at the futureGEO Community Workshop
HAM RADIO 2026, Friedrichshafen
Saturday, 27 June 2026
Hosted by AMSAT-DL e.V. with the ESA Satellite Communications Group
Author: Michelle Thompson W5NYV on behalf of ORI
A 501(c)(3) all-volunteer open-source nonprofit https://www.openresearch.institute
Version 1.0 — May 2026
1. EXECUTIVE SUMMARY
The amateur radio satellite community is asked, at HAM RADIO 2026, what should come after Qatar-OSCAR 100. AMSAT-DL and the ESA Satellite Communications Group have framed three concepts for discussion: an evolutionary bent-pipe successor (Enhanced QO-100+), a software-defined regenerative payload (Digital Innovation Lab), and an exploratory mm-wave pathfinder.
The Open Research Institute proposes that a regenerative, software-defined path is the best way forward. None of the concepts are a perfect alignment, but Concept 2 is the closest fit. The next step with regenerative software-defined hardware is the best next step, and that step can be taken now, with substantially reduced risk, because the necessary open-source building blocks exist and have been and are being demonstrated.
This paper describes those building blocks. It describes the Opulent Voice (OPV) air interface and the Haifuraiya (“High Flyer”) system architecture that uses it. It describes the
working demonstration hardware that ORI is bringing to the workshop: a complete FDMA uplink modem, a C++ reference modulator and demodulator, a human-radio user interface, the Pluto-side ARM application stack, and a hardware-verified DVB-S2 transmit chain suitable for the downlink. It explains why open-licensing this stack serves amateur radio better than any alternative on offer.
The claim made here is firm: an open, regenerative, voice-and- data unified architecture, built on the demonstrated Opulent Voice / Haifuraiya stack, best serves the international amateur radio community in geostationary orbit.
2. WHAT QO-100 TAUGHT US
QO-100, Es’hail-2’s narrowband and wideband amateur transponders, has been an extraordinary success. It proved that geostationary amateur access is technically and operationally viable, that a bent-pipe transponder remains a workhorse, and that even a modest amateur payload, well placed, can transform the daily life of thousands of operators.
QO-100 also taught us where the limits lie:
- A bent-pipe transponder cannot perform onboard processing. Spectrum is shared, not allocated. Misbehavior on the uplink lands directly on the downlink. There is no system discipline beyond operator etiquette.
- Modulation and access are constrained by what each operator can transmit. The transponder cannot impose framing, FEC, or scheduling.
- Bandwidth is finite, and so is downlink EIRP. Many users share a fixed pipe. Quality degrades with offered load.
- Voice quality, when digital voice is used at all, depends on legacy low-bitrate codecs designed for narrowband VHF repeaters. Codec 2 at 3.2 kbps, AMBE at 3.4 kbps. These voice encoders deliver quality that is technically a very long way behind what mobile telephony has offered for over two decades.
- Modes are siloed. Switching from voice to data, or voice to image, means stopping, retuning, and starting again on a separate scheme.
A successor payload that simply does more of the same will inherit all of these constraints. A regenerative, software-defined payload removes them.
3. THE OPULENT VOICE PROTOCOL
Opulent Voice is ORI’s open-source, patent-free digital voice and data protocol for amateur radio. The current specification is OPV Protocol v1.1 (January 2026). All parameters in this section are taken verbatim from the protocol specification and the published reference implementations.
3.1 Design Philosophy
OPV is built on five principles:
- Voice always wins. Voice transmission has absolute priority over all other traffic types, with carefully bounded exceptions for authentication.
- Modern codec quality. Opus is used at a minimum of 16 kbps, delivering audio quality that exceeds amateur expectations set by Codec 2 (3.2 kbps) and AMBE (3.4 kbps).
- Unified protocol. One frame format carries voice, text, file transfer, control, and authentication. There is no mode switching. Priority queues and UDP port assignment handle multiplexing.
- Open source. Every block of the protocol is fully documented, patent-free, and available under internationally recognized open licenses.
- 40 ms frame timing. The reference implementation aligns the air interface to the 40 ms Opus audio callback, so one voice packet exactly fills one over-the-air frame.
3.2 Physical Layer Parameters
Modulation MSK (Minimum Shift Keying)
Symbol rate 54,200 baud
Frequency deviation ±13,550 Hz
Frame duration 40 ms
Frame length 2168 symbols
Sync word 0x02B8DB (24 bits)
Channel coding Rate-1/2, K=7 convolutional
G1 = 0x4F (171 octal)
G2 = 0x6D (133 octal)
Interleaver 67 × 32 block with bit reversal
Randomizer CCSDS 8-bit LFSR,
polynomial x^8 + x^7 + x^5 + x^3 + 1
The 24-bit sync word 0x02B8DB was selected by exhaustive search of all 24-bit sequences. Of the 6,864 sequences that achieve an 8:1 peak-to-secondary-lobe ratio in the soft correlation output, 0x02B8DB is one, and it is mnemonic: “Oh to be eight dB!”
3.3 Frame Structure
The pre-encoding OPV frame is 134 bytes, structured as:
| Offset | Size | Field |
| 0 – 5 | 6 bytes | Station ID (Base-40 callsign encoding) |
| 6 -8 | 3 bytes | Authentication token |
| 9 – 11 | 3 bytes | reserved |
| 12 – 133 | 122 bytes | voice, control, text, data payload |
After randomization, rate-1/2 convolutional encoding, and 67×32 block interleaving, the payload occupies 2144 bits. The 24-bit sync word brings the on-air frame to 2168 symbols, or exactly 40 ms at 54,200 baud.
3.4 Higher Layers
Above the physical layer, OPV uses standard internet conventions:
- IP/UDP for transport.
- RTP for Opus voice payloads, de-jitter, and station ID hash.
- COBS (Consistent Overhead Byte Stuffing) for framing variable-length data that may span multiple OPV frames.
- Per-frame Base-40-encoded station identification.
- Per-frame authentication token enabling controlled access without proprietary encryption.
This is the protocol stack used in ORI’s terrestrial point-to- point Opulent Voice demonstrations and in the Haifuraiya satellite reference design.
4. THE DEMONSTRATION HARDWARE
ORI is bringing the following working, open-source designs to the futureGEO Community Workshop. All are public repositories under github.com/OpenResearchInstitute, and all are inspectable, buildable, and modifiable today.
4.1 pluto_msk — The Uplink Modem (VHDL / FPGA)
The pluto_msk repository is an Opulent Voice MSK modem targeting the AMD Zynq Z-7020 SoC inside the Analog Devices “PLUTO clone” SDRS. The modem and supporting blocks come from ORI’s open RTL library:
- msk_modulator
- msk_demodulator
- nco (numerically controlled oscillator)
- pi_controller (proportional-integral loop)
- prbs (pseudo-random bit sequence generator)
- power_detector
- lowpass_ema (exponential moving average filter)
The current design targets the LibreSDR platform (Zynq 7020 with LVDS interface to the AD9363 transceiver).
License: CERN-OHL-W (Open Hardware Licence, weakly reciprocal).
4.2 opv-cxx-demod — C++ Reference Modulator and Demodulator
A self-contained C++ implementation of both the OPV modulator and demodulator, designed to interoperate with the FPGA modem and with Interlocutor. It runs on a host computer alongside PlutoSDR or LibreSDR hardware. It is the reference for the air-interface, and verifies that the FPGA implementation matches the protocol specification bit-for-bit.
License: GPLv3.
4.3 dialogus — Pluto ARM/Linux Application
C-code applications that run on the ADALM-Pluto’s ARM cores inside the Zynq processing system. Dialogus is the device-side companion to Interlocutor, handling IIO operations, frame header construction, FEC, interleaving, randomization, MSK register configuration, statistics, and UDP encapsulation. It is the production-grade glue that makes the FPGA modem and the user interface work together.
License: GPLv3.
4.4 Interlocutor — Human-Radio Interface
A Python-based human-radio interface for Opulent Voice. It provides a web GUI, voice input and output, real-time text chat, slash-command extensions for operator workflow, and an audio device management layer. It is the user-facing terminal that turns the FPGA modem and the Pluto application into an actual amateur radio station. It has been deployed in full- duplex interoperability tests between PlutoSDR and LibreSDR.
License: GPLv3.
4.5 dvb_fpga — DVB-S2 Downlink Encoder
The dvb_fpga repository implements RTL components for a full DVB-S2 transmitter. All six core blocks are verified in simulation, in hardware testing, and the design is currently used for Q0-100 stations in the EU:
- Baseband scrambler
- BCH encoder
- Bit interleaver
- Constellation mapper
- LDPC encoder
- Physical-layer framing
Supported configurations:
- Frame types: Normal and Short
- Constellations: 8PSK, 16APSK, 32APSK
- Code rates: 1/4, 1/3, 2/5, 1/2, 3/5, 2/3, 3/4, 4/5,
5/6, 8/9, 9/10 - Reference build: Zynq UltraScale+ xczu4cg, 300 MHz target
- Per-frame parameter switching with no reset required
Behavior is validated to match GNU Radio’s DVB-S2 reference output. License: CERN-OHL-W.
This block is exactly what a Haifuraiya-style regenerative payload needs on its downlink: a re-modulator that takes multiplexed amateur traffic and emits a single, standards- compliant, ground-receivable DVB-S2/S2X signal.
4.6 What This Adds Up To
Considered together, these repositories constitute a complete, open-source, FPGA-and-software amateur radio communications system:
5. HAIFURAIYA: THE SYSTEM ARCHITECTURE
Haifuraiya, or “High Flyer”, is ORI’s reference geostationary mission concept. The architecture is described in the ORI public documentation set and summarized here from the OPV multiplexing reference document.
Many single-channel FDMA uplinks from ground-station user terminals, each one an Opulent Voice signal as described in Section 3, are received by the satellite. Each uplink is demodulated on board. Demodulated OPV frames carry, in their headers, the transmitting station’s Base-40 callsign and an authentication token. The payload determines whether the satellite forwards the frame, in accordance with the AAAAA protocol (access, authentication, authorization, accounting, auditing). Accepted frames are multiplexed onto a single DVB-S2/S2X downlink.
This is a regenerative payload. The satellite does not relay RF; it processes packets. That has significant consequences.
- The uplink and downlink are decoupled. Misbehavior on one uplink does not corrupt the downlink for everyone else. The satellite can mute, rate-limit, or drop bad actors at
the source. - Spectrum is used efficiently. FDMA uplinks at modest baud rate aggregate cleanly onto a wide downlink at one of the high-rate DVB-S2 modcods.
- The downlink is receivable with standard, inexpensive, widely available consumer DVB-S2 hardware. Every amateur on the planet has access to receive equipment.
- Per-frame authentication enables coordinated access without closing the system. Amateur openness is preserved; abuse is bounded.
- The system can carry voice, text, files, telemetry, and IoT traffic in one unified protocol, exactly as concept 1 in the workshop call envisions. Except we do this natively, without bolting features on top of a bent pipe.
Haifuraiya is not a bent-pipe-plus. It is a different class of mission, and it is closest to the class of mission concept 2 of the workshop description points toward.
6. MAPPING TO THE THREE WORKSHOP CONCEPTS
The workshop call presents three concepts. ORI’s Haifuraiya proposal relates to each as follows.
Concept 1 — Enhanced QO-100+ (Evolutionary)
The Enhanced QO-100+ concept proposes “classic bent-pipe narrowband and wideband transponders, an advanced beacon architecture, multi-band downlinks and additional functions such as text and image transmission, e.g. for emergency and disaster communication, IoT.”
Every “additional function” listed is a native, designed-in capability of Opulent Voice. Voice, text, image data, control, and IoT-class telemetry share one frame format with priority queuing. Where Concept 1 must add features to a bent pipe, Haifuraiya already has them.
Concept 2 — Digital Innovation Lab (Regenerative SDR)
This is the natural home of the Haifuraiya / Opulent Voice architecture. The workshop call notes a single risk for this concept: “the risk of being very software-heavy.” Section 8 below addresses that risk directly. The short form: the heavy lifting is in FPGA RTL, not in CPU software, and the RTL is already implemented, simulated, and verified on hardware.
Concept 3 — High Frequency Pathfinder (mm-wave)
ORI does not propose Haifuraiya as a replacement for an mm-wave pathfinder. The two are complementary. An open digital baseband stack like Opulent Voice / dvb_fpga is, in fact, exactly what a pathfinder needs in its supporting telemetry, beacon, and experimental data paths.
7. WHY OPEN SOURCE BEST SERVES AMATEUR RADIO
7.1 Patent-Free, License-Friendly
The Opulent Voice voice codec is Opus, a royalty-free, patent- free standard developed by the IETF. There are no per-radio license fees, no proprietary vocoder ICs, no export-controlled algorithms, no surprise patent claims. By contrast, every commercial amateur digital voice system in widespread use today depends on proprietary, royalty-bearing vocoders that cannot be freely implemented, freely shared, or freely modified.
7.2 Inspectable and Auditable
Every block in the Haifuraiya stack, the physical layer modulation, FEC, interleaving, randomization, framing, multiplexing, downlink encoder, is published source. Any amateur, any institution, and any regulator can verify exactly what is happening on the air. Open inspection is a prerequisite for amateur radio in jurisdictions that prohibit unlicensed encryption.
7.3 Distributable and Modifiable
CERN-OHL-W (used by pluto_msk and dvb_fpga) and GPLv3 (used by Interlocutor, Dialogus, and opv-cxx-demod) are recognized open licenses with strong reciprocity. Anyone can fork, port, improve, or replace any block. Anyone, anywhere can compile the bitstream. Different SDR platforms can host the design. Standards committees can incorporate it. None of this requires asking permission from a vendor.
7.4 Educational
Open hardware and open software are the most powerful teaching tools amateur radio possesses. Students can read the VHDL, run the simulations, watch the constellation, change a parameter, and learn. A closed-source satellite payload teaches nothing beyond its user interface.
7.5 Long-Lived
Open designs outlive their original maintainers. The CERN-OHL and GPL ecosystems are demonstrably durable. A bent-pipe analog payload is, ultimately, a single point of failure. An open, regenerative payload’s reference design is preserved indefinitely, can be re-flown, and can be extended.
8. ADDRESSING THE “SOFTWARE-HEAVY” RISK
The workshop description identifies a single risk for the Digital Innovation Lab concept: “But with the risk of being very software-heavy.” This concern deserves a direct response.
8.1 The Heavy Lifting Is in FPGA, Not CPU
The OPV modem, the demodulator, the FEC, the interleaver, the multiplexer, and the downlink DVB-S2 encoder all live in FPGA fabric. They are deterministic, real-time, and timing-closed. They are not running as Python or C on a general-purpose CPU. This is not software in the sense the workshop call is worried about; it is hardware design expressed in HDL. Demodulation and decoding done in software is done with validated and verified routines in dedicated processors, such as the A53 quad core on the ZCU102.
8.2 The FPGA Designs Exist and Have Been Built
Every block named in this paper has been synthesized, place- and-routed, and run on real silicon. The pluto_msk modem has been built for both PlutoSDR (Zynq 7010) and LibreSDR (Zynq 7020 with LVDS). The dvb_fpga blocks are verified in both simulation and hardware testing. We are not asking the amateur satellite community to fund a research program. We are asking it to adopt designs that already work.
8.3 The Air Interface Has Been Flown Terrestrially
Full-duplex Opulent Voice interoperability has been demonstrated end-to-end between Interlocutor with opv-cxx-demod on PlutoSDR, and Dialogus with the FPGA modem (Locutus) on LibreSDR. The protocol works. The implementation works. What remains is space qualification, not invention.
8.4 The Operating System Layer Is Small
The Pluto-side ARM application is C, not a heavy framework. The user terminal is Python, not a stack of microservices. It was deliberately designed to be internet-free, HTML5/CSS/Javascript only, and accessible by design. The complexity budget is dominated by the deterministic FPGA path. Software is a complement, not a foundation.
9. WHAT ORI PROPOSES
ORI proposes that the futureGEO Community Workshop:
- Adopt regenerative onboard processing as the technical baseline for the next amateur radio geostationary payload.
- Adopt Opulent Voice as the uplink air-interface reference, in coordination with the IEEE P1954 standards effort.
- Adopt DVB-S2/S2X as the downlink, using the open-source dvb_fpga implementation as the reference encoder.
- Adopt the Haifuraiya FDMA-uplink / multiplexed-downlink architecture as the system reference.
- Treat the demonstration hardware presented at HAM RADIO 2026 as the starting point for a working group, not as a finished product, with a written collaboration framework
between participating organizations.
ORI commits, for its part, to continued open development, public availability of every block under CERN-OHL-W or GPLv3, formal written partnership agreements with collaborators, and ongoing standards-track contributions through IEEE P1954.
10. CONCLUSION
Qatar-OSCAR 100 is a landmark. The next amateur radio geostationary payload should be a landmark of a different kind: the first regenerative, fully open-source, voice-and-data unified amateur GEO mission in history. The technology to build it is on the table at HAM RADIO 2026. The licensing is permissive. The protocols are documented. The hardware works.
This is what best serves amateur radio.
APPENDIX A — OPULENT VOICE TECHNICAL PARAMETERS
Modulation MSK
Symbol rate 54,200 baud
Frequency deviation +/- 13,550 Hz
Frame duration 40 ms
Frame length 2168 symbols
Sync word 0x02B8DB (24 bits)
Sync PSLR 8:1 (soft correlation)
Channel code Rate-1/2, K=7 convolutional
G1 = 0x4F (171 octal)
G2 = 0x6D (133 octal)
Interleaver 67 × 32 block with bit reversal
Randomizer CCSDS 8-bit LFSR x^8 + x^7 + x^5 + x^3 + 1
Payload (pre-encoding) 134 bytes
Station ID 6 bytes (Base-40)
Authentication token 3 bytes
Reserved 3 bytes
Voice / data 122 bytes
Voice codec Opus, ≥ 16 kbps
Transport IP / UDP, RTP for voice
Framing COBS
License (RTL) CERN-OHL-W
License (software) GPLv3
APPENDIX B — REPOSITORIES
pluto_msk github.com/OpenResearchInstitute/pluto_msk
opv-cxx-demod github.com/OpenResearchInstitute/opv-cxx-demod
dialogus github.com/OpenResearchInstitute/dialogus
interlocutor github.com/OpenResearchInstitute/interlocutor
dvb_fpga github.com/OpenResearchInstitute/dvb_fpga
Protocol spec Inside interlocutor:
opulent_voice_protocol.md
Related standards work IEEE P1954